this case happened by accident when I run model train with Pycharm IDE
but it must appear when I run on Linux terminal with python xxx.py
is it a version inconsistency or something else?
this case happened by accident when I run model train with Pycharm IDE
but it must appear when I run on Linux terminal with python xxx.py
is it a version inconsistency or something else?
It seems OpenMP
fails with this error. If you are seeing the same issue in the latest PyTorch release, could you create a GitHub issue so that we can track and debug it, please?
hi, ptrblck!
my torch version is ‘1.12.0’
Additionally, I adopt python’s Multiprocess class to implement the parallel model training over different gpu device.
I don’t know whether there exist conflict between torch’s OpenMP and the Multiprocess